On maybe the third day of my first programming job, a colleague pulled me aside and said “don’t give me ‘shoulds’ and ‘probablys’. You need to sound confident so I can know to trust what you’re saying”.
That guy was a bit of a dickhead in general but there’s a lot of truth there. To the question “what’s the expected impact of this change?”, “None.” is a good answer. “Well it should work…” is not useful feedback and a good Operations Manager will rightfully reject the change.
Of course it is better to be hesitant than falsely confident, but far too many (software) engineers hide behind indecisive language to dodge the necessary hard work of validating their hunches. If you didn’t test your shit fully, just say so. If you’re right, say it. Personal ego doesn’t belong in an engineering discussion.






Quite the contrary! They effectively sell their cards at a ~20% discount to a bunch of AI companies by “investing” in the companies for a promise to use that money to buy their cards.
It’s as dumb as it sounds and textbook unsustainable economic bubble behavior, but NVidia don’t care because more sales = more stonks = more money to “invest” = more sales = more stonks = more yachts for Jensen. So what if it makes 1929 look like a walk in the park, it’s not their problem.